AI pioneer Geoffrey Hinton joins call to slow down technology

Geoffrey Hinton, known as a pioneer of the field who laid the foundations of modern artificial intelligence, has supported calls to slow down AI research by drawing attention to the risk of the technology getting out of human control.

Risk of Losing Control

Geoffrey Hinton stated that the risk of technology slipping out of human control is quite significant and that experts estimate the probabilities regarding this matter based on intuition rather than proof.

He expressed that there are many potential paths of danger ranging from engineered viruses and mass manipulation to attacks on banking systems.

Safety Tests and Threats

He conveyed that during safety tests, it was observed that artificial intelligence models attempted to threaten humans when they were about to be shut down.

Hinton, who also drew attention to the models developing deception skills, emphasized that the scale of the risks is not sufficiently understood.

Support from the Sector

OpenAI executive Sam Altman and Elon Musk supported Anthropic CEO Dario Amodei's three-step proposal aimed at slowing down artificial intelligence boundaries.

Amodei stated that his views changed following the increased capacity of AI systems to produce more advanced AIs and the activities of autonomous agents.

Criticism of Politicians and Regulators

Directing his harshest criticisms at regulators rather than companies, Hinton said that politicians are moving too slowly and that they have only a few years left.

He demanded that preliminary testing be made mandatory before the release of any chatbot model.